Open

Category Archives: Tech

Tamara News  /  Blog  /  Tech
01 Sep 2026
AI chip interconnect bottleneck

The Real AI Chip Bottleneck Isn’t the Chip — It’s the Wires

Categories Tech

The AI chip interconnect bottleneck took center stage as SEMICON Taiwan 2026 opened in Taipei on 31 August. The event drew more than 100,000 semiconductor professionals from 65 countries to International Semiconductor Week. SEMI’s Terry Tsao delivered the conference’s central message, and it was blunt. Moving data between AI chips can now consume more energy than the computation itself. That shifts the industry’s hardest problem from individual chip performance to system-level architecture.

Why the AI chip interconnect bottleneck changes the conversation

For years, chipmakers competed primarily on transistor density and raw compute throughput. SEMICON Taiwan’s 2026 agenda reflects a shift toward a different metric entirely. The new focus is how efficiently data moves between thousands of chips working together inside a single AI training cluster. AI models have grown, and so has the number of chips that must communicate constantly during training. The wiring and interconnect standards linking them have not kept pace with compute gains. That gap creates exactly the kind of bottleneck Tsao described.

AI chip interconnect bottleneck

Who is racing to solve the wiring problem

Nvidia has staked much of its roadmap on faster interconnect standards to keep pace with the chips it sells. Its GPU Technology Conference earlier this year saw Jensen Huang announce roughly $1 trillion in expected orders for its Blackwell and Vera Rubin chip generations through 2027. Those orders depend partly on solving the same data-movement problem SEMICON Taiwan highlighted. Component suppliers focused on optical and copper interconnect technology have seen renewed investor interest as a result. Faster, more energy-efficient connections between chips could unlock compute gains that raw chip design alone cannot deliver.

The regulatory backdrop shaping who can compete

The interconnect race is unfolding alongside tightening US export controls on advanced AI chips and the semiconductors that power them. The Commerce Department has signaled further regulatory action on chips and AI is coming. The Department has separately signed letters of intent worth $874 million with seven companies to strengthen the domestic compute supply chain. Those two threads are tighter export rules abroad and new incentives at home. Together they are pushing US-based chip and interconnect makers to treat domestic manufacturing capacity as a competitive necessity, not a cost center.

Taiwan’s own chipmakers face a delicate balance in this environment. They supply both US and Chinese customers. Interconnect standards adopted at events like SEMICON Taiwan increasingly carry geopolitical weight alongside their technical merits. TSMC and other Taiwanese suppliers have avoided taking public positions on which standard should win out. They prefer to support multiple approaches until the market settles on a default, a stance that lets them keep selling into both American and Chinese supply chains without picking a side.

How this affects AI training timelines industry-wide

Data-center operators say interconnect limits already stretch some training runs longer than the chips themselves would require. A cluster with thousands of top-tier chips gains little if those chips spend significant time waiting on data transfers rather than computing. That waiting time translates directly into higher costs. Labs still pay for chip time regardless of utilization. Several cloud providers have begun advertising interconnect specifications alongside raw chip counts for the first time this year. Industry analysts say the shift reflects customer demand for clearer efficiency metrics. Buyers increasingly ask about bandwidth between chips before they ask about raw chip counts, according to several vendors present at the conference.

Smaller AI labs without the scale to negotiate custom interconnect solutions face a particular disadvantage. They typically rely on off-the-shelf networking gear that larger labs have already moved past. That gap can leave them paying similar chip costs for meaningfully worse effective performance. Several smaller labs have started pooling compute purchases specifically to negotiate better interconnect terms, mimicking a strategy large cloud providers pioneered years earlier. Industry groups say more such consortiums are likely to form if interconnect costs keep rising faster than raw chip prices.

What happens next

International Semiconductor Week runs through the rest of the week in Taipei. More technical sessions are expected to detail specific interconnect standards vying to become the industry default. Whichever approach gains traction will shape how quickly AI labs can actually deploy the chips they are paying for. That includes labs racing to fund new compute capacity, like DeepSeek.

Frequently Asked Questions

What is the AI chip interconnect bottleneck?

It refers to the energy and speed limits of moving data between AI chips working together in a cluster, which industry leaders now say can consume more energy than the chips’ own computation.

What is SEMICON Taiwan 2026?

It is International Semiconductor Week’s flagship event, held in Taipei starting 31 August 2026 and drawing more than 100,000 semiconductor professionals from 65 countries.

Why does this matter for Nvidia and its rivals?

Nvidia’s roadmap depends on faster interconnect standards to support roughly $1 trillion in expected orders for its Blackwell and Vera Rubin chips through 2027.

How do US export controls factor in?

Tightening restrictions on advanced AI chips are pushing chipmakers to treat domestic manufacturing and interconnect capacity as strategically necessary, alongside new federal incentives for the compute supply chain.

Related Coverage

Sources

01 Sep 2026
DeepSeek funding round

Inside the DeepSeek Funding Round That Could Precede a 2027 IPO

Categories Tech

A DeepSeek funding round worth roughly 50 billion yuan, about $7.4 billion, is nearing a close. Sources cited by China Money Network put the pre-money valuation near $74 billion. DeepSeek is the Chinese AI lab that rattled US markets in January 2026. Its model matched Western systems at a fraction of the training cost. DeepSeek is using this raise to fund about 1 gigawatt of new compute capacity, ahead of a planned initial public offering.

What the DeepSeek funding round is actually paying for

The bulk of the new capital is earmarked for compute infrastructure, not research headcount. That reflects how much of the competitive pressure in frontier AI now comes from access to chips and data-center capacity, rather than algorithmic breakthroughs alone. DeepSeek is racing rivals including Alibaba’s Qwen, Tencent’s Hy4, and Zhipu’s GLM for both compute and engineering talent. Sources familiar with the round say the financing is intended to close by the end of August. That would position DeepSeek to file for an IPO before the end of 2026, with a public debut targeted for 2027 on Shanghai’s Star Market.

DeepSeek funding round

Image credit: Feliciagrace-bytesrack / Wikimedia Commons (CC BY 4.0)

Why this funding round matters beyond DeepSeek

DeepSeek’s January 2026 model release briefly wiped more than $500 billion from AI-linked US equities. It demonstrated that competitive AI systems no longer required the enormous training budgets US labs had assumed were necessary. This new funding round suggests the opposite lesson is now taking hold inside DeepSeek itself. Staying competitive at the frontier still requires massive, sustained capital, even for a lab that built its reputation on efficiency. That shift has implications for every AI lab claiming a cost advantage. It implies efficiency gains buy time rather than permanently lower capital requirements.

The chip-supply backdrop shaping the round

DeepSeek’s fundraising lands against a backdrop of tightening US export controls on advanced AI chips. The Commerce Department has signaled further regulatory action on chips and AI is coming. Chinese AI labs have had to route around restrictions on the most advanced Nvidia and AMD chips. A war chest of this size gives DeepSeek more room to secure whatever compute capacity remains available to it. That capacity could come through domestic chipmakers or through channels that comply with existing export rules.

Analysts tracking China’s AI sector say the round also signals confidence from domestic investors that Beijing will keep backing frontier AI development despite the external pressure. That confidence matters as much as the capital itself, since it shapes whether other Chinese labs can raise on similar terms.

How investors are pricing DeepSeek against its global rivals

A $74 billion valuation puts DeepSeek well below OpenAI and Anthropic on paper. It still ranks among the most valuable AI labs outside the United States. Private investors backing the round are betting that DeepSeek can keep its cost advantage even as it spends more on infrastructure. That bet carries real risk. Training costs across the industry have climbed as models grow larger. DeepSeek’s efficiency edge could narrow if rivals adopt similar techniques.

Domestic Chinese funds make up a large share of the round, according to people familiar with the terms. That matters for a specific reason. It reduces DeepSeek’s exposure to foreign capital controls and export-linked investment restrictions. Those restrictions have complicated fundraising for other Chinese tech companies this year.

Some investors involved in the round declined to comment publicly, citing the sensitivity of ongoing negotiations. That reticence is typical for pre-IPO rounds of this size, where terms often shift until the final signing. Bankers advising on the deal have also stayed quiet about the exact investor list. That silence is standard practice ahead of a Shanghai listing, where regulators scrutinize pre-IPO ownership disclosures closely and expect confidentiality until filings become public.

What happens next

Investors will be watching whether the round closes at the reported terms by DeepSeek’s end-of-August target. They will also watch whether an IPO filing follows on the timeline sources have described. A successful Shanghai listing would give Chinese investors direct exposure to one of the country’s most closely watched AI companies. It would also test whether public markets value DeepSeek’s efficiency reputation as highly as private investors currently do.

Frequently Asked Questions

How much is DeepSeek raising in this funding round?

Reports put the round at roughly 50 billion yuan, about $7.4 billion, at a pre-money valuation near $74 billion.

What will DeepSeek use the money for?

The company plans to use most of the capital to build roughly 1 gigawatt of new compute capacity, according to sources familiar with the round.

Is DeepSeek planning an IPO?

Yes. Sources say DeepSeek could file for an IPO by the end of 2026, with a public debut targeted for 2027 on Shanghai’s Star Market.

How does this relate to DeepSeek's January 2026 model release?

That release showed DeepSeek could match Western AI systems at a fraction of the training cost, briefly erasing over $500 billion from AI-linked US stocks. This funding round suggests staying competitive still requires large, sustained capital investment.

Related Coverage

Sources

30 Aug 2026
WhatsApp OTP API PHP guide cover showing a POST send request

The PHP Signup Step That Stops Codes Going Missing

Categories Tech

You do not need an SDK or a WhatsApp Business Platform build to verify a phone number. This guide walks a working WhatsApp OTP API PHP flow using cURL: send the code, verify what the user submitted, and branch on the errors that matter.

This guide uses Replio’s WhatsApp OTP endpoint because it is a single JSON POST with no SDK to install. The shape of the flow is the same whichever provider you use, so the structure transfers.

WhatsApp OTP API PHP flow from send through verify
The four steps of a WhatsApp OTP API PHP integration.

Before you write any PHP code

Three things need to exist first. A WhatsApp number connected to your provider account. At least one approved Authentication template on the WhatsApp Business Account. An API key.

Keep the key server-side. It sends from your verified business number, so a leaked key means someone else messaging your customers under your brand.

WhatsApp OTP API PHP: sending the code

One POST sends the code. You can pass a code you generated yourself, or omit it and let the provider generate, hash and store one for you.

$ch = curl_init("https://engine-production-2647.up.railway.app/api/otp/send");
curl_setopt_array($ch, [
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_POST           => true,
  CURLOPT_HTTPHEADER     => [
    "Authorization: Bearer " . getenv("REPLIO_OTP_KEY"),
    "Content-Type: application/json",
  ],
  CURLOPT_POSTFIELDS => json_encode([
    "phone"           => "447911123456",
    "idempotency_key" => "signup-" . $signupId,
  ]),
]);

$body   = curl_exec($ch);
$status = curl_getinfo($ch, CURLINFO_HTTP_CODE);
curl_close($ch);

$data = json_decode($body, true);
if ($status >= 400) { throw new RuntimeException($data["detail"]["code"]); }

A success looks like this:

{
  "ok": true,
  "sent_to": "+447911123456",
  "template": "verify_code",
  "credits_charged": 1,
  "verify_enabled": false
}

Read verify_enabled carefully. It is true only when you omitted the code. That flag tells you whether the verify endpoint has anything to check.

Verifying what the user typed back

If you let the provider generate the code, check what the user typed with a second call.

$ch = curl_init("https://engine-production-2647.up.railway.app/api/otp/verify");
curl_setopt_array($ch, [
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_POST           => true,
  CURLOPT_HTTPHEADER     => [
    "Authorization: Bearer " . getenv("REPLIO_OTP_KEY"),
    "Content-Type: application/json",
  ],
  CURLOPT_POSTFIELDS => json_encode([
    "phone" => "447911123456",
    "code"  => $userInput,
  ]),
]);

$body   = curl_exec($ch);
$status = curl_getinfo($ch, CURLINFO_HTTP_CODE);
curl_close($ch);
$data = json_decode($body, true);

if ($status < 400 && !empty($data["verified"])) { return completeSignup(); }

switch ($data["detail"]["code"]) {
  case "incorrect_code":    return show("That code is not right.");
  case "code_expired":      return show("That code expired. Send a new one.");
  case "too_many_attempts": return forceResend();
}

A correct code returns { "ok": true, "verified": true }. A wrong or expired one is a normal 400, not a 200 with a false flag. Handle it as an error branch.

Where this fits in your signup flow

Treat the send and the verify as two separate states in your own model, not one blocking call. Send the code, store the signup attempt, and return control to the user. Verify runs later, when they submit the form.

That separation matters when things go wrong. If the verify call fails, you still hold the signup attempt and can offer a resend without losing the user’s progress. If you couple the two, a network blip drops them back to the start.

Rate limits sit on the recipient as well as the account. Five codes to one number per hour, and ten verify attempts per number per ten minutes. Surface a clear message when you hit those rather than a generic failure, because a user who resends four times in a minute will hit them.

One detail trips people up on the first run. The phone number goes in international format as digits. A leading plus sign, spaces and dashes are accepted and stripped, but a local-format number without a country code is rejected as invalid_phone. Normalise before you send.

Only a delivered send costs a credit. Rejected requests, rate limits and test-mode calls are free, so strict validation on your side costs nothing.

Handling the errors that actually happen

Branch on the machine-readable code field, never on the human-readable message. The message wording can change at any time; the codes are the contract.

  • incorrect_code — wrong digits. Let the user retry.
  • code_expired — past its time to live. Offer a resend.
  • too_many_attempts — five wrong guesses burn the code. Force a new one.
  • recipient_rate_limited — five codes to one number in an hour. Back off.
  • upstream_error — WhatsApp was unreachable. Safe to retry.

Two habits that save you money and credits

Pass an idempotency key. Networks time out after a send has already happened, and a blind retry sends a second code and spends a second credit. With a key tied to the signup attempt, a retry returns the original result instead.

Then build against a test key. A test credential validates the whole request and applies every rule, but sends nothing and bills nothing.

Hardening the flow before launch

Set a short time to live. Five minutes is the common default and it limits the window for a stolen code. Cap wrong guesses. Never log the code itself.

If you are weighing this against your current SMS provider, we compared the two channels in WhatsApp OTP vs SMS. The full parameter list and error table live in the Replio WhatsApp OTP API reference, and Meta documents the template rules in its message template guide. For context on running WhatsApp as a support channel too, see our piece on answering WhatsApp and Instagram without working nights.

Frequently asked questions

Do I need an SDK for PHP?

No. It is one JSON POST with a bearer token, so your language’s standard HTTP client is enough.

Should I generate the code myself?

Either works. Pass your own code and the provider only delivers it. Omit it and the provider generates one, stores a hash, and gives you a verify endpoint.

How long does a code stay valid?

Five minutes by default, configurable between 60 and 1800 seconds when the provider generates the code.

What if the user never receives it?

Check the error code on the send. If the number has no WhatsApp account the send fails, which is your cue to fall back to SMS.

Is the code stored anywhere?

Replio stores only a sha256 hash of codes it generates, never the code itself. Codes you supply are not stored at all.

Related Guides

29 Aug 2026
SpaceX insider lockup expiration

$101 Billion in SpaceX Stock Just Became Sellable — Here’s the Catch

Categories Tech

SpaceX insiders got their first chance to cash out since the company’s June 2026 IPO. The SpaceX insider lockup expiration freed up to 911.5 million shares for potential sale starting August 6. That stock is worth an estimated $101 billion. The stock touched a new low around the release date. Investors were weighing how much of that stock would actually hit the market.

Unlike most IPOs, SpaceX did not use a single lockup expiration date. It built a staggered schedule instead. That spreads the risk of a selling wave across several months.

SpaceX insider lockup expiration

What the SpaceX insider lockup expiration actually released

The first tranche unlocked August 6, 2026. It let insiders sell up to 20% of their eligible shares. That is as much as 911.5 million shares, or roughly $101 billion at prevailing prices. A second tranche of 319 million shares was scheduled to unlock August 12. More releases continue through the rest of the year. Not every insider share became sellable. A separate block of up to 455.8 million shares stays locked. SpaceX stock has been trading below its $135 IPO price, and that condition was built into the lockup terms.

CEO Elon Musk and a group of other major shareholders operate under a separate, extended lockup. It does not expire until June 2027. So the bulk of the shares now eligible for sale belong to earlier employees and outside investors, not Musk himself.

Why the stock hit a new low around the release

Markets typically price in some selling pressure ahead of a lockup expiration. Existing shareholders now have the option to realize gains, or cut losses, after months of being unable to trade. SpaceX shares fell to a new low around the August 6 date. That fits investors positioning for a supply increase, regardless of how much stock insiders actually chose to sell.

The staggered structure aimed to limit exactly this kind of shock. It spreads eligible sales over multiple dates instead of releasing everything at once. Even so, the first tranche alone, roughly $101 billion in newly sellable stock, was large enough to move the share price.

What’s still locked and for how long

The full 180-day lockup period runs through early December 2026. At that point, up to 5.33 billion shares in total become eligible for trading. Additional tranches will unlock on a rolling basis between now and then. The SpaceX insider lockup expiration is better understood as an ongoing process through year-end, not a single event that already ended on August 6.

Staggered lockups like this one are becoming more common among large, high-profile IPOs. Underwriters use them to avoid dumping an entire float onto the market on a single day, which can overwhelm demand and send a stock sharply lower in a short window. SpaceX’s schedule spreads that risk across roughly six months instead.

What to watch through the rest of the unlock schedule

Investors will watch each tranche for one signal: how much stock insiders actually sell versus hold. That behavior shows how confident early shareholders feel about SpaceX’s valuation going forward. The December milestone matters most. That is when the full 180-day lockup ends. It also marks when the largest pool of shares becomes eligible, well before Musk’s own extended lockup expires in mid-2027.

How SpaceX’s valuation holds up through each unlock will also shape sentiment around other recent high-profile IPOs using similar staggered structures. A smooth series of releases would support the case for spreading lockups over months. A sharp drop at any single tranche would raise fresh questions about whether staggering actually reduces the shock, or simply delays it. The next scheduled unlock date will be an early signal either way.

Frequently asked questions

How much SpaceX stock became sellable after the lockup expired?

Up to 911.5 million insider shares became eligible for sale starting August 6, 2026. That is worth roughly $101 billion, when the first tranche of SpaceX’s post-IPO lockup expired.

Did all SpaceX insiders get to sell their shares?

No. A separate tranche of up to 455.8 million shares stays locked because SpaceX stock trades below its $135 IPO price. Musk’s own shares fall under an extended lockup until June 2027.

What is the full SpaceX insider lockup expiration schedule?

SpaceX used a staggered schedule instead of one release date. 20% of eligible shares unlocked August 6, another 319 million were set for August 12, and the full lockup runs through early December 2026.

How many SpaceX shares could eventually be tradable?

Up to 5.33 billion shares become eligible for trading once the full 180-day lockup ends in early December 2026. Not all of that stock will necessarily be sold.

Did SpaceX’s stock price drop after the lockup expired?

Reports describe the stock hitting a new low around the lockup expiration. That reflects investor concern that a wave of insider selling would pressure the share price.

Why did SpaceX structure its lockup differently from typical IPOs?

A staggered release schedule spreads potential selling pressure across multiple dates instead of one event. The goal is generally to reduce the shock of a single massive sell-off.

Related coverage on Tamara News

Sources

29 Aug 2026
tech layoffs wave 2026

127,000 Tech Jobs Gone in 2026 — And the Companies Cutting Aren’t Struggling

Categories Tech

Four more household names cut jobs in August 2026. That pushed this year’s tech layoffs wave 2026 past 127,000 positions across 281 companies. The total already tops all of 2025, with a third of the year still to run. Apple, TikTok, LinkedIn and Netflix each announced cuts within weeks of each other. None of them are actually shrinking their business overall.

Companies posting solid results while cutting staff is the defining feature of this year’s layoff cycle. August’s announcements make the pattern harder to ignore.

tech layoffs wave 2026

Who cut jobs in the latest tech layoffs wave

Apple laid off more than 60 employees from its Vision Pro team, plus additional staff from its Siri division. Both units have struggled. The Vision Pro headset has failed to find a mass audience since its 2025 launch. Siri’s promised AI overhaul has faced repeated delays. Apple rarely announces layoffs at all. That is part of why this round drew outsized attention.

TikTok cut 75 positions, concentrated in its TikTok Shop and global e-commerce divisions. The company called the move necessary given recent restructuring. LinkedIn is cutting roughly 875 employees, about 5% of its workforce. The reorganization aims at faster-growing areas of the business. Netflix is closing two gaming studios: Night School Studio in Los Angeles and Moonloot in Helsinki. The goal is to refocus its gaming unit on titles for kids, parties and mainstream players, not narrative games.

The scale of the tech layoffs wave 2026 so far

Total tech job cuts for 2026 reached 127,180 across 281 companies by late August, according to industry tracking. That is nearly 5,000 more than the full-year total for 2025, with four months still left. This trajectory puts 2026 on pace to be the heaviest year for tech layoffs since the sector’s post-pandemic correction. Most of the companies doing the cutting are not in financial distress.

Some sectors within tech are affected more than others. Consumer hardware and gaming units, like Apple’s Vision Pro team and Netflix’s closed studios, have seen cuts tied to weak product performance. Enterprise and social platforms, like LinkedIn and TikTok, describe their cuts as strategic reallocation instead. That split matters for anyone trying to read the headline number as a single trend rather than two distinct stories running in parallel.

Why profitable companies keep cutting staff

The pattern across Apple, TikTok, LinkedIn and Netflix looks like reallocation, not retrenchment. Each company is investing heavily in AI infrastructure, AI product features, or both. Each has framed its cuts as freeing up budget and headcount for growth areas, not a response to falling revenue. LinkedIn explicitly called its cuts a reorganization toward growing parts of the business, not a downturn response.

That framing matters for how workers and investors read these announcements. A company cutting staff because a product failed, like Apple’s Vision Pro team, differs from one cutting staff to fund a strategic pivot, like LinkedIn’s reorganization. Both still show up in the same layoff count.

What the rest of 2026 could bring

Four months remain in the year, and the total already exceeds 2025’s full count. Tracking sites expect the number to keep climbing. The fourth quarter typically brings additional rounds as companies finalize next year’s budgets. A further wave tied to 2027 planning is plausible before the year closes. Whether it concentrates in struggling product lines, like Apple’s Vision Pro and Siri teams, or in strategic reorganizations, like LinkedIn’s, will shape how the labor market absorbs the cuts.

Workers laid off from struggling product teams and workers laid off from reorganizing but profitable teams often land very differently in the job market. The first group typically competes for a shrinking pool of similar roles. The second group frequently gets recruited quickly by companies expanding in the same growth areas their old employer is chasing.

Frequently asked questions

How many tech jobs have been cut in 2026?

Tech job losses reached 127,180 across 281 companies by late August 2026. That is nearly 5,000 more than the total for all of 2025, per tracking cited by Fast Company and Yahoo Finance.

Which companies announced cuts in the latest tech layoffs wave?

Apple, TikTok, LinkedIn and Netflix all announced job cuts in August 2026. That list already included Meta, Microsoft, Oracle and Samsung earlier in the year.

Why is Apple cutting jobs on its Vision Pro team?

Apple laid off more than 60 employees from its Vision Pro unit, plus staff from its Siri team. The headset has sold slowly, and Siri’s AI overhaul has faced delays.

How many employees is LinkedIn cutting?

LinkedIn is cutting about 5% of its workforce, roughly 875 employees. The reorganization aims to focus staff on faster-growing business areas.

Why is Netflix closing game studios during the tech layoffs wave?

Netflix is shutting Night School Studio in Los Angeles and Moonloot in Helsinki. It wants to refocus its gaming division on titles for kids, parties and mainstream audiences.

Is the tech layoffs wave 2026 driven by AI investment?

Many affected companies are increasing spending on AI infrastructure and AI roles at the same time. A share of the cuts reflects reallocation toward AI priorities, not a broad-based downturn.

Related coverage on Tamara News

Sources

1 … 21 22 23 24 25 26