THE APEX TIMES
Google orders more than 3 million TPUs from Intel for 2028, report says
The reported deal would be a major boost for Intel’s contract-manufacturing push as demand for AI accelerators strains rivals’ chip supply.
Alphabet’s Google has placed an order with Intel to manufacture more than three million Tensor Processing Units, or TPUs, for delivery in 2028, according to The Information as reported by Reuters on Monday, June 8, 2026.
TPUs are custom chips designed by Google to speed up machine learning workloads by performing large matrix operations efficiently. Google makes TPUs available through its cloud infrastructure, including Cloud TPU deployments, where they are used to train and run AI models at scale. The reported order underscores how central custom silicon has become to Google’s AI computing strategy.
The Reuters report said Intel declined to comment on the claim. Alphabet (Google’s parent) and Nvidia also did not immediately respond to requests for comment, and Reuters said it could not independently verify the report.
The scale and timing of the order matter to Intel because they point to continued customer interest in Intel as an alternative manufacturing partner for advanced AI chips. Reuters tied the interest to a broader industry problem, in which fast-rising AI chip demand has left Taiwan Semiconductor Manufacturing Company, or TSMC, struggling to secure adequate capacity, prompting some chip designers to look for additional suppliers.
Intel has been working to rebuild its foundry business after years of setbacks in advanced chip manufacturing. In recent materials, Intel describes Intel Foundry as a “systems foundry” model that includes manufacturing process nodes plus advanced packaging and assembly and test services, an approach aimed at producing “systems of chips” built from multiple dies, which is increasingly common in AI accelerator designs.
Alongside the Google report, Reuters noted that Nvidia is also evaluating whether Intel’s technology could be used to manufacture a processor that combines four graphics chips into a single unit, though Nvidia has not placed an order. Intel shares rose more than 9% in early trading on the reporting, as investors weighed fresh evidence of momentum in the company’s turnaround under CEO Lip-Bu Tan.
Several key commercial details remain unclear. The reports did not disclose the contract value, the specific TPU generation involved, the manufacturing process node, or whether the work would lean primarily on wafer fabrication, advanced packaging, or both. Because Reuters said it could not verify The Information’s reporting independently, the timing and exact scope of any eventual agreement are still uncertain.
Why It Matters
- If confirmed, the TPU volume would represent a significant external validation of Intel Foundry’s ability to deliver advanced AI compute components, not just internal Intel products.
- A meaningful shift away from single-supplier concentration could reduce delivery risk for AI chip designers as global manufacturing capacity remains tight.
- The report suggests Google’s custom silicon strategy is also a supply-chain strategy, using more than one manufacturing ecosystem to meet future demand.
- Intel’s ability to attract large AI chip customers could influence how quickly the market re-rates Intel’s foundry turnaround prospects.
Sources
- story (Yahoo Finance RSS link)
- Reuters via UOL (reporting the The Information claim)
- Reuters (reporting the same claim)
- Google Cloud documentation: Introduction to Cloud TPU (what TPUs are)
- Intel Foundry fact sheet PDF (what Intel Foundry includes, packaging and systems-of-chips framing)
- Google blog explainer: What is a TPU (additional plain-English context)
- Image
Key Facts
- Google placed an order with Intel to manufacture more than three million TPUs for delivery in 2028, according to The Information and reported by Reuters on June 8, 2026.
- TPUs are Google’s custom ASICs used to accelerate machine learning workloads, and are offered through Google’s Cloud TPU infrastructure.
- Intel declined to comment on the report, while Alphabet and Nvidia did not immediately respond to requests for comment.
- Reuters said it could not independently verify The Information’s claim.
- The reported order would bolster Intel’s contract-manufacturing business as AI demand strains chip capacity at suppliers such as TSMC.
- Intel positions its foundry offering as a “systems of chips” approach that combines process technology with advanced packaging and assembly and test services.
Technology Related
Anthropic rolls out “Claude Duet” phone agent that proactively handles email and shopping, upping pressure on rivals like Meta and OpenAI
The new assistant is positioned as a more proactive consumer “agent,” moving from chat toward tasks such as reading messages and making purchases. The move lands in a category Meta has also been working, including efforts the company has described publicly as reaching millions of users.
SEEQC says NVIDIA-linked decoding approach improves quantum error correction performance
In a new set of results, the chip-based quantum computing company SEEQC reported performance gains tied to a digital interface-optimized decoder architecture built around NVIDIA Ising and NVQLink, aiming to make quantum error correction more practical.
Nvidia’s DSX POD enters the conversation as executives weigh how fast AI gets deployed
In a Wall Street panel discussion tied to Nvidia’s latest DSX POD announcement, investors and analysts focused on one question: what happens when AI moves beyond pilots and into day-to-day enterprise workflows.
Rockwell Automation adds NVIDIA Nemotron support to FactoryTalk Conductor
The industrial automation company says its FactoryTalk Conductor platform will recognize NVIDIA’s Nemotron open models, aiming to bring more configurable AI capabilities into industrial operations workflows.
Apple pushes iPhone Duo developer updates live this Friday, expanding apps across fold positions
Apple says apps optimized for its first foldable iPhone will roll out starting Friday, highlighted by seamless transitions between the outer and inner displays, new iOS tooling for folding layouts, and early feature work from Netflix, Microsoft and major productivity, media and games partners.
QumulusAI boosts closed-lease funding tied to NVIDIA Blackwell GPUs, adding $123.6 million via Tech Finance
The neocloud infrastructure provider says the expanded equipment lease financing with Tech Finance Co., LLC will support additional NVIDIA Blackwell graphics processing units without issuing new equity.
Yahoo Finance asks investors whether to trim or add Intel amid ongoing market debate
A new Yahoo Finance piece frames the latest question facing Intel shareholders, weighing the case for adding versus trimming, but does not offer new Intel disclosures in the post itself.
Oracle’s “wildcard” funding mix targets Microsoft’s cloud war, but the plan’s staying power is unclear
A market report argues Oracle is using a high-variance strategy to bankroll its push against Microsoft in cloud and AI spending, mixing stock sales, new debt, and customer prepayments. Whether that financing holds together could shape who gains control of the next wave of enterprise computing budgets.
Yahoo Finance report says Microsoft’s AI data-center spending could create a massive competitive moat through 2030
A new analysis frames Microsoft’s accelerating investment in AI infrastructure, cloud growth, and shareholder returns as a potential long-term advantage, while acknowledging that the company has not laid out a fully quantified “moat” roadmap.
Google-backed virtual power plant signs 1 gigawatt “bring your own capacity” deal in PJM
The agreement with energy platform Voltus is aimed at pulling distributed resources into PJM’s grid, with a stated goal of supporting energy affordability and delivering financial value to participating homes and businesses.