Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsThe seven companies in this 2024 watchlist were Anthropic, Cohere, AI21 Labs, Hugging Face, Aleph Alpha, Scale AI and Tenstorrent. They were not the seven largest AI businesses, nor a ranking of the best investments. The list mixed model makers with a developer platform, data infrastructure and a chip designer—companies with different customers, risks and routes to growth.
This is a historical look at the companies selected as emerging challengers and infrastructure providers entering 2024. “Watch” means strategically significant, not guaranteed to succeed or available to buy as a public stock. The original list appeared on December 26, 2023, and used editorial signals such as company age, backing and management experience rather than a published scoring system. AI Business’s original list is useful context, but those signals do not prove revenue, product-market fit or durable advantage.
What “top AI companies” meant in this 2024 list
The companies occupied different parts of the AI stack. Anthropic, Cohere and AI21 Labs developed language models; Hugging Face supported model discovery and development; Aleph Alpha paired model work with sovereign-AI positioning; Scale AI supplied data and evaluation services; and Tenstorrent designed processors and licensed semiconductor IP. They were not interchangeable competitors.
The selection is best understood as a watchlist of potential challengers and strategically important suppliers, not a league table by revenue, valuation, market share or model performance. It also did not attempt to list every central AI company: large technology firms controlled substantial compute, cloud distribution and capital. Epoch AI’s October 2024 analysis estimated that Google, Microsoft, Meta and Amazon collectively held computing capacity equivalent to hundreds of thousands of NVIDIA H100 GPUs, while OpenAI and Anthropic relied substantially on rented or partner infrastructure. Epoch AI’s computing-capacity analysis illustrates the structural advantage incumbents had; it does not establish that any one company would win.
#1 Best Overall
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
Seven AI companies to watch in 2024
1. Anthropic: frontier models and Claude
Anthropic develops the Claude family of general-purpose AI models and was positioned as a direct rival to OpenAI. Its safety and reliability focus, along with access to enterprise distribution through Amazon Web Services, made it a prominent model company to watch.
Amazon said in March 2024 that its total investment in Anthropic had reached $4 billion. The companies also linked Anthropic’s model development to AWS infrastructure, including Trainium and Inferentia chips, and made Claude available through Amazon Bedrock. Amazon’s announcement describes the investment and partnership from Amazon’s perspective. In November 2024, Anthropic said Claude was available to tens of thousands of companies through Bedrock and described work with AWS on Trainium accelerators and the Neuron software stack. Anthropic’s announcement is a company-reported account of that reach, not an independent measure of adoption or customer outcomes.
The partnership gives Anthropic capital, infrastructure and a route to AWS customers, while making the company dependent on a major cloud partner. Amazon benefits from offering Claude to customers, but strategic investment and distribution are not proof of market leadership. Anthropic also faces high frontier-model costs and competition from OpenAI, Google, Meta and open models.
2. Cohere: language models aimed at businesses
Cohere builds language models and generative-AI tools with an enterprise-first orientation. Founded in 2019 by Aidan Gomez and Nick Frosst, both associated with Google Brain, it sought to help organizations integrate language capabilities into their products and workflows rather than center its story on a consumer chatbot.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →The original 2023 watchlist cited relationships involving firms such as Spotify, Glean, Oracle and McKinsey, as well as backing from investors including Oracle, Salesforce and NVIDIA. Those names indicate potential access and interest, not necessarily paid production deployments, revenue scale or customer satisfaction. Cohere’s business case rests on customizable business applications and enterprise deployment needs; its challenge is that cloud providers and other model companies can offer competing services directly, while open models may put pressure on API pricing.
For product and platform information, see Cohere’s official site. Availability and pricing depend on product and deployment; no single current price is established here.
3. AI21 Labs: language models and NLP services
Founded in 2017, AI21 Labs developed language models including the Jurassic family and offered developer access through AI21 Studio. Its background in natural-language processing and enterprise-facing APIs made it a longer-running startup in a market that rapidly became crowded after generative AI went mainstream.
Rank #2
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
The 2023 article named enterprise relationships involving Capgemini, Samsung, Ubisoft, AWS, Google Cloud, Snowflake and Dataiku. Partnerships and named customers are not the same as proof of broad paid deployment. AI21’s opportunity was to serve specialized language and business use cases; its risk was that larger model providers could compete on quality, price and distribution, while public visibility remained lower than that of the best-known chatbot brands.
Its official site and AI21 Studio describe its products. Model performance and API economics can change, so comparisons need to specify the model, task and date.
4. Hugging Face: the open-source AI ecosystem
Hugging Face differs from the model labs: its importance comes primarily from its platform and developer community. Founded in 2016, it hosts models, datasets and machine-learning tools that let developers discover, share, test and deploy work from many organizations. The 2023 list highlighted its support for projects such as Llama 2, partnerships with firms including Dell and AWS, and its own model releases.
A broad repository can create network effects and make Hugging Face a useful distribution layer without giving it control of the strongest proprietary models. Its ecosystem influence also does not automatically translate into high-margin revenue. For organizations evaluating a hosted model, the repository alone is not a safety or compliance review: check the model card, license, dataset documentation, provenance and maintenance status before deployment. Licenses differ, and “open” does not always mean unrestricted use or fully open-source.
Explore the Hugging Face platform and its pricing information; plan availability and costs should be verified for the specific service and date.
Free tools Windows power users keep installed
One-click scans. No signup required.
5. Aleph Alpha: European and sovereign-AI positioning
Germany-based Aleph Alpha developed the Luminous model family and positioned its AI offering for enterprise and government use, including organizations concerned with control over sensitive information and deployment conditions. Founded in 2019, it attracted backing that the original article associated with Bosch Ventures, Hewlett Packard Enterprise, SAP and companies linked to Schwarz Group.
Its European base and sovereign-AI positioning could appeal to public-sector and regulated buyers seeking alternatives to relying wholly on U.S. providers. But sovereignty messaging is not, by itself, proof of legal sovereignty, technical control or superior model quality. Buyers need to assess where data is processed, which providers operate the infrastructure, contractual terms and applicable rules. Aleph Alpha also faced the scale and distribution disadvantages common to smaller model providers, alongside complex public-sector procurement.
Rank #3
- EVOLUTION RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
See Aleph Alpha’s official site for its current offering; enterprise terms may depend on deployment and contract.
6. Scale AI: data, evaluation and AI infrastructure
Scale AI provides data-labeling, data-management, evaluation and AI development services. Founded in 2016, it occupies an enabling layer: models need curated data and evaluation, and organizations building AI systems need ways to assess and improve them. The 2023 article cited work involving OpenAI, Meta, Microsoft, Toyota and General Motors, as well as commercial and government customers.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchServing multiple model developers can make a data provider less dependent on one model brand. Yet labeling and quality control can be labor-intensive, and customers may automate more work, bring it in-house or use synthetic data. Data provenance, privacy and labor practices are material considerations. Customer names and partnerships do not establish revenue, profitability or the scale of any one engagement.
See Scale AI and its solutions overview for current services. The offering is oriented toward organizational needs rather than a simple consumer subscription.
7. Tenstorrent: AI processors and semiconductor IP
Tenstorrent designs AI processors and licenses AI and CPU intellectual property, placing it in hardware rather than the language-model business. Founded in 2016, it drew attention in part because CEO Jim Keller had a long semiconductor career, and the 2023 article named investors including Samsung Catalyst Fund, Hyundai Motor Group, Kia, Fidelity Ventures and Maverick Capital.
Its potential is tied to demand for alternatives in AI training and inference, including customized silicon. But a chip is only one part of a deployed system: customers need manufacturing capacity, compatible software, developer support, reliable performance at scale and a compelling total cost for real workloads. NVIDIA’s position rests not only on processors but also on software, networking and ecosystem adoption. NVIDIA reported fiscal 2024 data-center revenue of $47.5 billion, up 217% from fiscal 2023; this scale helps explain the hurdle, but does not directly compare Tenstorrent products. NVIDIA’s SEC filing gives the company’s fiscal-year results.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Tenstorrent’s possible advantage is not established by leadership credentials or a stated hardware ambition alone; workload-level benchmarks, availability, software maturity and customer deployments would matter. See Tenstorrent’s site for current product and company information.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How the seven companies compare
| Company | AI layer | Primary relevance | Potential advantage | Key risk |
|---|---|---|---|---|
| Anthropic | Foundation models | Claude for developers and enterprises | Model development and AWS distribution | Compute costs and cloud-partner dependence |
| Cohere | Language models | Business integrations and enterprise use | Enterprise-first product focus | Hyperscaler competition and pricing pressure |
| AI21 Labs | Language models and NLP APIs | Developer and enterprise language services | NLP experience and API orientation | Competition and changing model economics |
| Hugging Face | Developer platform and ecosystem | Model, dataset and tool discovery | Community and cross-provider reach | Monetization, licensing and repository variability |
| Aleph Alpha | Models and enterprise AI | European public sector and regulated buyers | Sovereign-AI positioning | Scale, procurement pace and global competition |
| Scale AI | Data and evaluation infrastructure | Organizations building and assessing AI | Services across model providers | Labor costs, provenance and automation |
| Tenstorrent | Processors and semiconductor IP | AI infrastructure teams and chip developers | Potential hardware alternatives and licensing | Manufacturing, software ecosystem and adoption |
The categories describe each company’s main role in this 2024 watchlist, not an exhaustive account of every product. Their public-company status is not presented as a current investment screen: the original list concerned emerging firms, and private-market access is not the same as ordinary public-stock availability.
What investors and business readers should verify
A notable investor, a prominent executive or a partnership can signal access to capital, infrastructure or customers. None establishes a durable business. For a more disciplined assessment, examine evidence by category rather than comparing unlike companies as if they were all chatbot vendors.
- Technical differentiation: Is the distinction tied to a specific model, chip, data capability or platform, and is the evidence dated and relevant to real workloads?
- Commercial traction: Are there disclosed recurring deployments or revenue, or only customer names, pilots and partnerships?
- Infrastructure access: Does the company have compute, cloud distribution, manufacturing or other scarce capacity—and how dependent is it on a partner?
- Ecosystem and defensibility: Are developers, integrations, proprietary data, switching costs or software tools likely to reinforce adoption?
- Economics and execution: Can the business manage compute expense, sales cycles, labor, capital needs and customer concentration?
- Risk exposure: Consider regulation, data rights, privacy, safety, licensing and the possibility that incumbents or open alternatives absorb the product’s features.
Enterprise buyers should add practical checks: security controls, data handling, deployment options, service commitments, integration work and total cost. A model that performs well in a demonstration may not meet a buyer’s production, compliance or reliability requirements.
Why partnerships can help—and complicate—the outlook
AI startups often need a large cloud provider, software distributor, investor or manufacturer to reach customers and obtain infrastructure. Anthropic’s AWS relationship shows the two-sided bargain: Amazon invested and offered cloud capacity and distribution, while AWS customers gained access to Claude. A strategic investor can also operate a competing platform. Partnership, investment, distribution and acquisition are distinct arrangements; none should be treated as proof of commercial success or control.
The same structural pressure applies across the list. Frontier-model firms face enormous compute requirements; enterprise providers compete with the platforms already used by corporate buyers; open-source ecosystems must find sustainable business models; data suppliers face automation and provenance questions; and chip designers must build both hardware and software adoption. These are reasons to watch the companies, not assurances that they will prevail.
What this list can—and cannot—tell a reader
The seven companies offer a map of different opportunities in the 2024 AI supply chain: models at Anthropic, Cohere and AI21 Labs; developer infrastructure at Hugging Face; sovereign-oriented enterprise AI at Aleph Alpha; data services at Scale AI; and hardware at Tenstorrent. A company may matter strategically without being the largest, most profitable or most visible AI business.
This is not a stock recommendation or a current ranking. Private-company exposure may be unavailable or unsuitable for ordinary investors, and the historical watchlist does not establish present-day performance. For a company decision, evaluate the specific product and contract; for an investment decision, seek current financial and ownership information and account for the risks of private markets.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
