TSMC has completed development and validation of its A16 process platform built with Super Power Rail, or SPR, a backside power delivery network that ETNews said preserves compatibility with existing design ecosystems. The report says TSMC is the first foundry to bring backside power supply technology to the Angstrom-class era while keeping current design infrastructure largely intact, a point that matters for large chip customers already invested in the N2 and N2P ecosystem. According to the report, the move addresses a growing bottleneck in sub-2 nm chip design, where power and signal routing compete for limited front-side space, raising routing congestion and voltage drop risks. TSMC’s SPR architecture shifts power delivery to the back of the chip and uses dedicated backside contacts to feed each transistor’s source and drain, while leaving front-side gate structures, cell size, and layout area nearly unchanged. ETNews said A16 can deliver 8% to 10% higher performance at the same power, or cut power use by 15% to 20% at the same speed, with chip density also rising 8% to 10% versus N2P. The report also said NVIDIA’s next-generation AI accelerator platform, Feynman, will use A16 and aims for mass production in the second half of 2028, alongside large-scale use of CPO and SoIC 3D stacking.

Samsung has pushed back its 1.4nm mass-production target from 2027 to 2029, extending its focus on the 2nm platform for at least three more years. The shift puts more weight on yield improvement, process maturity and customer adoption, rather than rushing into another node transition. The article argues that Samsung is still developing its 2nm family rather than standing still. In addition to standard SF2, the company is working on derivative 2nm processes for different use cases and plans to introduce backside power delivery in more advanced versions. Samsung has also taken a cautious stance on High-NA EUV, with Vice President Park Chang-min saying the technology still needs to mature and may only become essential for production at A10 and below, or the 1nm class and beyond. That caution stands out as TSMC keeps advancing. According to the article, TSMC’s N2 has already entered mass production, and A16, its first post-2nm node, is scheduled for mass production in the second half of 2026. A16 will use TSMC’s SPR backside power delivery technology. TSMC says that versus N2P, A16 can deliver an 8% to 10% speed gain at the same power, a 15% to 20% power reduction at the same speed, and up to about 10% higher chip density. The piece frames the latest roadmap shift as a sign that leading-edge competition is no longer just a race to the next number. Cost, yield, power delivery, packaging and ecosystem execution are becoming just as important as transistor scaling itself.

Wall Street is changing how it values listed bitcoin miners that are moving into AI and high-performance computing. The market is shifting away from hash rate, bitcoin output and BTC holdings toward energized megawatts, signed IT load and delivery execution. Core Scientific’s latest quarter shows the new model in action: high-density colocation revenue reached $136.7 million, or about 83% of total revenue, while self-mining revenue fell 66% year over year. TeraWulf and Hut 8 have also signed multibillion-dollar data center leases, but much of that capacity will not be delivered until 2027 or 2028. VanEck says the industry’s biggest challenge is execution, not demand, and estimates a near-term funding gap of about $50 billion across the sector.

A group of publicly listed Bitcoin miners moving into AI and high-performance computing is being judged by a new set of metrics. Investors are no longer focused only on hash rate, Bitcoin production, and BTC held on balance sheets. They are asking how much power a company controls, how much of that capacity is already energized or backed by clear grid interconnection arrangements, how much has been leased to AI customers, and how much is already delivered and billing. The shift is showing up in company results and contract announcements. Core Scientific said its high-density hosting business generated $136.7 million in revenue in the second quarter of 2026, about 83% of total revenue, while self-mining revenue fell roughly 66% year over year to $21.54 million. TeraWulf signed a 20-year data center lease with Anthropic covering about 401 MW of critical IT load and carrying an expected value of about $19 billion, while Hut 8 announced an additional 352 MW lease in Texas valued at $9.8 billion. The opportunity is large, but delivery remains limited. Based on data through June 4, 2026, VanEck estimated that the companies involved had delivered only about 25% of their leased capacity and faced a near-term funding gap of about $50 billion. In this market, the most valuable asset is not planned gigawatts on paper, but power that can be energized on time, financed, built into high-density data centers, and rented under long-term contracts by reliable customers.

Bitcoin mining companies that have moved into artificial intelligence and high-performance computing are drawing richer valuations as pressure on mining economics continues. According to BlockBeats, falling bitcoin prices and weaker mining profitability have pushed more listed miners to look for steadier revenue streams through AI and HPC contracts. Over the past year, TerraWulf, IREN and Cipher Digital more than doubled in share price, while MARA Holdings, which shifted to AI later than peers, fell about 40% over the same period. Mining economics also deteriorated sharply: hashprice dropped from about $63 per PH/s in July last year to roughly $31.8 per PH/s, while Bitcoin network hashrate declined from 1.14 ZH/s to around 900 EH/s, a drop of about 21%. CoinShares said that, as of the first quarter of 2026, miners with AI/HPC contracts traded at an average enterprise value multiple of about 12.3x, compared with 5.9x for pure bitcoin miners. The industry had also signed roughly $70 billion in cumulative AI/HPC contracts. Riot Platforms added to that trend last week by signing a 20-year lease agreement with Anthropic valued at about $9.1 billion. CoinShares also said a recovery in pure bitcoin mining remains possible if bitcoin returns to its roughly $126,000 record high from last October.

Over the past year, leading listed Bitcoin miners have been recasting themselves as energy infrastructure platforms, AI cloud providers and digital infrastructure companies. The pivot has already reshaped revenue at Core Scientific and TeraWulf, where hosting and high-performance computing now account for most sales. But the article argues that the real cost of an all-in shift is flexibility: once miners replace ASIC fleets with GPU-based AI capacity and lock themselves into 15- to 20-year contracts, they may not be able to switch back when Bitcoin mining becomes profitable again. The piece points to Marathon Digital and Hut 8 as more cautious examples. Both kept mining operations and Bitcoin holdings in place while funding AI expansion, preserving an option that fully converted miners are giving up.

Listed bitcoin mining companies tracked in the report saw their combined realized hashrate fall from 368.3 EH/s in the fourth quarter of 2025 to 319.0 EH/s in the second quarter of 2026, a decline of 13.4% over six months, according to Bitcoin.com News as cited by Odaily. Over the same period, the Bitcoin network’s quarterly average hashrate dropped 10.6%. Excluding Bitdeer, the decline among the relevant miners widened to 21.2%. The figures also show a growing revenue split inside the sector. Core Scientific reported $136.7 million in hosting revenue in the second quarter, roughly five times its $27.5 million in bitcoin mining revenue, with hosting making up 83% of quarterly revenue. TeraWulf posted $31.9 million in HPC leasing revenue, accounting for 71% of total revenue and exceeding its $12.8 million in bitcoin mining revenue. Bitdeer moved in the opposite direction on hashrate, with realized hashrate up 44% from the fourth quarter of 2025 to 63.0 EH/s. In June, its self-mining and joint-mining hashrate stood at 73 EH/s and 15.9 EH/s, and it produced 990 BTC that month, up 388% year over year. Cango, by contrast, fell from 44.8 EH/s to an estimated 16.5 EH/s over the same six-month period.

Second-quarter earnings from listed crypto mining companies showed a widening gap between firms still relying on Bitcoin mining and those already booking meaningful revenue from AI and high-density data center operations. MARA posted a $611.3 million net loss, including a $343 million unrealized fair-value loss on Bitcoin, even as its quarterly production edged up to 2,422 BTC. Riot Platforms mined more Bitcoin as well, yet its mining revenue fell as the average Bitcoin price declined and network hashrate increased. At the same time, Core Scientific and TeraWulf reported that hosting and HPC leasing had become major contributors to current revenue rather than distant promises. Core Scientific generated $136.7 million in high-density hosting revenue, up from $10.6 million a year earlier, while TeraWulf said HPC leasing made up about 71% of its quarterly revenue. The results also show why headline contract values can mislead. Multi-billion-dollar agreements signed by miners are not the same as revenue already recognized in quarterly statements. What matters now is how much capacity has been delivered, how much rent is being booked, and whether those new businesses can cover construction, depreciation and financing costs.