• Skip to main content
  • Skip to secondary menu
  • Skip to footer

Technologies.org

Technology Trends: Follow the Money

  • Technology Events 2026-2027
  • Sponsored Post
  • Technology Markets
  • About
    • GDPR
  • Contact

PrismML, the Startup That Shrinks AI Models to Run on an iPhone, Is in Talks With Apple

July 15, 2026 By admin Leave a Comment

A small Caltech spinout called PrismML has done something that looked implausible a year ago: it compressed a 27-billion-parameter large language model down small enough to run entirely on an iPhone, and Apple is now evaluating the technology. PrismML CEO Babak Hassibi told CNBC that Apple and other companies have been measuring the startup’s models for speed, energy efficiency, and on-device performance. “They’re really evaluating our technology right now,” Hassibi said, characterizing the talks as very early but progressing. For Apple, whose entire AI strategy hinges on keeping processing on the device rather than in the cloud, the timing could hardly be more pointed.

What PrismML Actually Did

The demonstration that got Apple’s attention: PrismML took Alibaba’s open-source Qwen 3.6 model, which has 27 billion parameters and weighs roughly 54 gigabytes in standard precision, and shrank it to under 4 gigabytes, a compression ratio above 90%, and ran it on an iPhone 17 Pro. Crucially, the company claims no meaningful loss in performance, and says the compressed model can still handle complex chat, reasoning, fully autonomous agents, and software coding. The startup publicly released compressed versions of Qwen under an Apache 2.0 license, along with custom kernels for Apple’s Metal framework so the models run on iPhone and Mac hardware. By PrismML’s numbers, the compressed models use 10 to 15 times less memory, generate responses 6 to 8 times faster, and consume 3 to 6 times less energy than full-precision versions on existing hardware.

The Trick: One Bit Instead of Sixteen

The core method is extreme quantization. Where conventional models store each internal weight as a 16-bit floating-point number, PrismML reduces each value to just one of a handful of possibilities, using 1-bit or ternary architectures where every weight is simply -1, 0, or +1. Hassibi compared it to the chip industry’s move from 8-bit to 4-bit computing, but taken further. The reason this saves so dramatically on both memory and energy is intuitive: multiplying by 1 or 0 is trivial compared to full floating-point math, so the model needs far less storage and far less power to run. PrismML is careful to frame the achievement as mathematics rather than an AI breakthrough, the work comes out of years of neural-network compression research at Caltech, not a new model or a cleverer training recipe.

Why This Matters for Apple Specifically

Apple has been fighting a losing battle against a hard constraint: the most capable AI models are simply too big for a phone. The most advanced parts of Siri are still large enough that Apple runs them on Nvidia chips inside Google Cloud, exactly the cloud dependency Apple wants to escape. Apple’s own new on-device model, AFM 3 Core Advanced, has 20 billion parameters but uses a sparse architecture where only 1 to 4 billion are active at any moment, a workaround that limits capability. PrismML’s compressed Qwen, by contrast, keeps all 27 billion parameters active simultaneously while fitting in under 4GB. If Apple could run models that large and dense on-device, it could move demanding features, computational photography, video generation, health and fitness tools handling sensitive personal data, off the cloud entirely, improving both speed and privacy. As analyst Carolina Milanesi of Creative Strategies put it, the more you can do on-device, the better, especially for health and medication data users want kept private.

The Backers

PrismML emerged from stealth earlier this year as a spinout of the California Institute of Technology, co-founded by Babak Hassibi, a professor of electrical engineering, alongside other PhDs who did the underlying compression research. It raised a $16.25 million seed round backed by Khosla Ventures, OpenAI’s first venture investor, along with Cerberus Capital and Caltech itself. Vinod Khosla has publicly called the work a “mathematical breakthrough” that could shift AI away from data-center dominance toward efficient edge deployment. The company frames its ambitions well beyond phones, positioning the technology for laptops, robotics, wearables, and industrial edge devices, and says it eventually intends to compress even trillion-parameter models to run locally.

Insight: The Skeptic’s Case on Chip Demand

The most interesting question isn’t whether PrismML’s compression works, it’s what happens to chip demand if it does, and here the analysts urge caution. The instinctive read is that shrinking models means needing far fewer chips, a potential threat to the memory and datacenter-GPU buildout driving the entire AI trade. But Gil Luria of D.A. Davidson argues that’s the wrong conclusion. Compression doesn’t eliminate the need for processors and memory, he says, it relocates them: “You’re still going to need the GPU, and you’re still going to need the memory.” Moving AI onto hundreds of millions of individual phones can actually be less efficient than shared datacenter infrastructure, because chips sitting in a phone are idle most of the time, whereas datacenter GPUs run near-continuously across many users. In other words, on-device AI might shift where the silicon lives rather than reduce how much is needed, and could even increase total memory demand as every premium phone ships with more RAM to hold these models. That nuance matters for anyone reading this as bearish for memory suppliers.

Insight: The Claims Still Need Independent Proof

Every number in PrismML’s pitch, the 90%-plus compression, the “no performance loss,” the speed and energy multiples, currently comes from the startup itself. Extreme quantization normally degrades a model’s accuracy, sometimes severely, which is precisely why “compress it and lose nothing” is such a strong claim. The fact that PrismML open-sourced its models under Apache 2.0 helps, because it invites the research community to verify the performance independently rather than taking marketing figures on faith. And Apple’s willingness to sit at the table is itself a meaningful signal, a company that runs its own compression research wouldn’t bother evaluating an outside startup unless it saw a genuine gap between what its models deliver and what the hardware could theoretically support. But “Apple is evaluating” is not “Apple is partnering,” and definitely not “Apple is acquiring.” The talks are exploratory, with no agreement, timeline, or deployment confirmed, and there’s no guarantee they lead anywhere.

Insight: A Broader Rewiring of Where AI Runs

Step back and PrismML is one data point in a larger structural question the whole industry is circling: how much AI belongs in the cloud versus on the device. The cloud model has three well-known pain points, privacy (data that leaves the device can be intercepted or subpoenaed), cost (every query to a remote GPU costs money, multiplied across hundreds of millions of users), and latency (a round trip to a datacenter is slower than local inference, which matters enormously for voice assistants and camera features). Compression that genuinely preserves capability attacks all three at once. If techniques like PrismML’s mature, the competitive advantage in consumer AI could tilt toward whoever controls the device and its silicon, which is exactly the position Apple has spent years and billions building toward with its Neural Engine. The irony is that Apple may end up needing an outside startup’s math to finally unlock the on-device strategy its own hardware was designed for.

What to Watch Next

Three things will tell whether this becomes real. First, independent benchmarks: now that the compressed models are open-source, third-party researchers can confirm or puncture the “no performance loss” claim, and that verdict will matter more than any demo. Second, whether Apple’s exploratory talks convert into a partnership, an acquisition, or nothing, Apple has bought AI startups before, but evaluating is not committing. Third, the competitive response: if compression this aggressive holds up, expect the other frontier labs and device makers to move fast on their own edge strategies, and expect memory and chip analysts to start modeling what a shift from datacenter to device actually does to demand. For now, PrismML has cleared the hardest bar, a working iPhone demo of a model this size, but the distance between a compelling demo and a shipping product is exactly where most breakthroughs stall.

Filed Under: News

Reader Interactions

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Footer

Recent Posts

  • TerraFirma Raises $100M Series A to Turn Heavy Construction Equipment Into Robots
  • PrismML, the Startup That Shrinks AI Models to Run on an iPhone, Is in Talks With Apple
  • OpenAI’s First Device Will Be a Moveable, Screenless AI Companion Speaker
  • IBM’s 25% Stock Fall Is Beginning of the End for Old School Software Giant
  • Why a Six-Axis Robot Arm Is Staring at a Green-Headed Tanager
  • Industrial Robotics Meets the AI Boom: What Cobots at Trade Shows Are Really Selling
  • Microsoft Trims 5,500 Jobs to Defend a $190 Billion Capital Program
  • South Korea Commits $590 Billion to Double Its Memory Chip Capacity
  • HyperLight Closes $80M to Move TFLN From Lab to Foundry
  • Odyssey Raises $310M to Build World Models on AWS Trainium

Media Partners

  • Market Analysis
  • Cybersecurity Market
  • App Coding
Enterprise Money Is Leaving Old School IBM for AI Infrastructure Companies
Why EU Tech Is Falling Behind the US: A Structural Diagnosis, Not a Cultural One
The HyperLight Threat to Coherent and Lumentum Ends Where Indium Phosphide Begins
SpaceX IPO (SPCX): A $1.75 Trillion Valuation Built on Selling 4% of the Company to People Who Watch Rocket Launches
What a Trillion-Dollar Cloudflare Actually Requires
The Repricing and the Drain: How SpaceX, OpenAI, and Anthropic Rewire the Index
Quantum Computing Equities: Market Segment Memo
Quantum Computing Stocks Face Violent Selloff the Moment Markets Reopen Tuesday
The $2.6 Trillion Signal: What Gartner’s AI Spending Forecast Actually Tells You
The Productivity Is Already Here. The Bubble Narrative Is Not.
Cybersecurity Stocks Rally as IBM CEO Flags Cyber Fears as a Top Customer Priority
Trump Administration Launches “Gold Eagle” Federal Clearinghouse for AI Cyber Threat Sharing
JadePuffer: Researchers Document the First Fully Autonomous AI Ransomware Attack
Aikido Acquires Root for a Reported $70 Million to Patch Open Source Without Forcing Upgrades
The three-week freeze on Anthropic’s most capable models is over
Miasma Supply Chain Worm Jumps to Go and Now Executes Inside AI Coding Assistants
Two-Factor Authentication Bypass: Attackers Brute-Force 2FA Systems, Gaining Access to Enterprise Accounts
France’s Tchap Government Messaging Breach Signals Weak Oversight of Encrypted State Communications
OpenSSL CVE-2026-45447: Heap Use-After-Free in PKCS#7 Verification Enables S/MIME RCE, Discovered With AI
Microsoft Patch Tuesday June 2026: Record 200+ Vulnerabilities in Single Release, Three Pre-Disclosure Zero-Days
PixVerse Closes Series C Extension at $439 Million and Pivots From AI Video Into Games
DigitalOcean Launches AI-Native Cloud at Deploy 2026
Verdent Updates AI Platform to Function as a Full Engineering Team for Solo Builders
The Side Project App Is Not Dead. The Side Project App Business Is.
The App Monetization Landscape Has Changed and Most Teams Have Not Caught Up
Building Offline-First Mobile Apps Is Harder Than It Looks and Worth It
State Management in React Native Has Too Many Options and One Right Answer
Mobile Accessibility Is the Case Developers Keep Ignoring
Testing Mobile Apps at Scale Without Losing Your Mind
App Store Optimization in 2026 Is a Different Game Than It Was

Media Partners

  • Market Research Media
  • Technology Conferences
  • API Coding
Getty Images Kills the $3.7 Billion Shutterstock Merger Rather Than Sell the Editorial Business the UK Demanded
Fox’s $22B Roku Deal: 4.6x Sales, Paid in 1.5x Stock
Tuesday Open: AI Earnings Engine Holds the Line as Iran Overhang Fades to Noise
China’s U.S. Treasury Holdings: The Great Repositioning (2021–2025)
Infographic: Why the 2025 CIPA Data Proves the APS-C Renaissance is Real
How WiFi Changed Media
Canva Acquires Simtheory and Ortto to Build End-to-End Work Platform
Netflix Price Hikes, The Economics of Dominance in a Saturated Streaming Market
America’s Brands Keep Winning Even as America Itself Slips
Kioxia’s Storage Gambit: Flash Steps Into the AI Memory Hierarchy
RAISE Summit, July 8-9 2026, Paris
CJS Securities 26th Annual New Ideas Summer Conference, July 9, 2026, White Plains, NY
SEMICON West 2026, October 13–15, San Francisco
Deutsche Bank Technology Conference 2026, August, Dana Point
ECOC 2026, September 20–24, Málaga
Citi Global Technology Conference 2026, September, New York
Goldman Sachs Communacopia + Technology Conference 2026, September, San Francisco
InfoComm 2026, June 13–19, Las Vegas
EBMI 2026, June 17–18, Frankfurt
FPGA Conference Europe, June 30 – July 2, 2026, Munich
Why Private Domain Data Is the Real Key to AI That Actually Works
Orkes Raises $60M to Bring Production-Grade AI Orchestration to Enterprise Developers
Form.io Launches MCP Server and Agentic Coding Toolset for Governed Enterprise AI Development
Appdome Upgrades MobileBOT Defense With Identity-First Mobile API Protection
Five SDK Generators Compared: Speakeasy, Stainless, Fern, APIMatic, and OpenAPI Generator
API Monetization Models That Work and the Ones That Drive Developers Away
gRPC in Production: What the Documentation Doesn't Tell You
Event-Driven Architecture vs Request-Response: Choosing the Right Communication Pattern
The Business Case for Internal APIs That Most Engineering Leaders Ignore
Breaking Changes: How to Avoid Shipping Them and What to Do When You Must

Copyright © 2026 Technologies.org

Media Partners: Market Analysis · Market Research · Referently · Photography