DeepSeek: What lies beneath for the channel?
Who is looking forward to selling DeepSeek and what are the geopolitical red flags? CRN speaks to the channel after China’s AI-powered chatbot made a splash in US tech stocks
A re-evaluation of AI strategies and investments, disruption to datacentre building, affordable AI adoption for MSPs, and a huge opportunity around services - this is what has surfaced for channel partners since DeepSeek’s arrival.
Last week, the American tech stocks took a hit after the launch of the Chinese AI assistant caught the global IT scene off guard.
With market losses between $700bn ($568.7bn) and $1tn, heavily impacted by the 17 per cent share drop of AI chip giant Nvidia, the whale brand made waves across all layers of the channel.
While its LLM training cost has been praised ($6m compared to over $100m for GPT-4), other aspects such as its lack of neutrality regarding Chinese politics have been heavily criticised.
But is this a long-lasting revolution or a temporary disruption in the world of AI, and how exactly does it affect the channel?
To answer these questions, CRN sat down with channel experts from Cisilion, QBS Software, Core Technology Systems, and Pax8, who shared their thoughts on the Chinese technology’s real impacts.
CRN: Does the advent of DeepSeek represent an opportunity for the channel? Or will it disrupt the already existing business models?
Mini Biswas, AI and security specialist at Cisilion: The news of DeepSeek AI is massively disrupting existing business models, you can see stock shares plummeting for many big organisations.
DeepSeek AI is classed as a cost-effective and high-performing AI model, and this is naturally leading to a re-evaluation of AI strategies and investments in our present time.
However, there are plenty of opportunities this news brings.
There's competition. Competition drives innovation and keeps prices in check.
Having multiple players in the AI field means more choices for consumers.
DeepSeek's AI models are developed at a fraction of the cost compared to other leading models.
This will make advanced AI capabilities more accessible to a broader range of businesses, allowing smaller companies to leverage AI without the hefty price tag, and add competitive edge as they pass on these savings to their clients.
With recent news on DeepSeek's models being available on platforms like Azure AI Foundry and GitHub, developers and businesses can experiment and integrate these models into their workflows more easily.
This democratisation of AI technology could spur innovation and even more new business models.
Alex Tatham, strategic board advisor at QBS Software: There are a couple of impacts.
Firstly, it might well disrupt some of the datacentre building going on in the UK as planners will start to look at the costs more closely.
Previously, the shortage of H100 chips drove the price high as end-users struggled to come first in the queue. The rush might slow down a bit.
ChatGPT had yet to formulate a channel model although this is coming along with other AI vendors seeking to sell business corporate licenses for a controlled AI environment.
DeepSeek has launched a corporate license model of their own, and QBS are looking forward to selling it!
Katie Sloan, marketing director at Core Technology Systems: DeepSeek’s advancements could reshape AI capabilities, but disruption isn’t inherently negative - it’s a catalyst for innovation.
For the channel, this means an opportunity to evolve service models, rethink AI integration strategies, and support customers in navigating new technologies.
The key for IT providers and MSPs is to remain agile, positioning ourselves as trusted advisors who help businesses extract real value from emerging AI solutions rather than being overwhelmed by change.
Eric Mink, VP of AI adoption AMEA at Pax8: If you look at the overall AI adoption rate within MSPs and their end-customers, partners struggle to justify AI costs for their SMB customers. As a result, this could foster affordable AI adoption at scale.
I think this could also level the playing field between SMB and enterprise.
The Pax8 marketplace strategy is perfectly positioned to distribute affordable AI solutions at scale.
For the MSPs, the value is not in reselling AI solutions.
The benefit to MSPs and their customers is through the delivery of value-added AI services, custom integrations, and industry-specific solutions.
MSPs will become AI consultants for their clients.
CRN: Do you think this will have a long-term effect on AI hardware and services?
Mini Biswas: We are already seeing a shift pattern where the world wants even more efficiency, consolidation, lowered costs and more Innovation.
Newer AI models are being designed with this in mind, this shift of having optimisation requiring less computational power and memory will follow which, may, eventually at some point in the future see a decline in the market on AI hardware.
This is why DeepSeek AI is making headlines, they managed to develop their AI model at a fraction of the cost compared to other leading models and that is bound to have an effect.
Regarding services - lowered costs, increased competition, fostering innovation and growth in AI will be the opportunity for the channel to welcome.
This is the value-add: humanising the technology, creating use cases, adding more context into prompting, and supporting the adoption for many businesses moving into a GenAI world.
Alex Tatham: Nvidia will still dominate this environment for a while.
DeepSeek are using older Nvidia chips and newer GPUs will still be in great demand.
Katie Sloan: Yes, DeepSeek’s impact could be significant, particularly in how AI models are trained and deployed.
If it drives efficiency improvements, we may see shifts in AI hardware requirements, potentially reducing demand for traditional high-power computing infrastructures while accelerating investment in new architectures.
From a services perspective, AI-driven automation will likely continue to redefine the skills and expertise required within IT teams, reinforcing the need for MSPs to offer strategic guidance on implementation, optimisation, and governance.
Eric Mink: We have seen what this news did in the short-term with the biggest drop in stock value ever recorded.
Watching recent stock market performance, we see there is massive tension in the market.
Big technology companies are all expected to rework their pricing models as a result.
We already see that every device being shipped is equipped with an NPU to process AI workloads locally meaning AI workloads no longer need to be processed in the cloud.
If DeepSeek’s model is really lightweight and efficient as stated in the press, MSPs and SMBs could run AI-powered Copilots locally, eliminating API costs.
This could significantly impact hyperscaler AI pricing.
Cloud AI is not going away as I see it: I expect it to look more like a hybrid AI model with local AI for simple tasks and cloud AI for complex tasks.
This will become the standard in which large-scale AI workloads (e.g., training, advanced analytics, and multi-device sync) still need cloud processing.
CRN: How are competing vendors likely to respond?
Mini Biswas: Competing vendors are likely exploring what’s next for them. Can they accelerate development to release their next advanced version of models sooner to market or look to collaborate with other AI models?
A few days back, OpenAI CEO Sam Altman and Microsoft CEO Satya Nadella hinted at major upcoming developments in their partnership in a recent post on X: “next phase of the Microsoft x OpenAI partnership is going to be much better than anyone is ready for".
We also heard recently that OpenAI has raised concerns of intellectual property that DeepSeek may have copied its work.
OpenAI claims DeepSeek used data produced by OpenAI's technology to improve their own AI models.
This would be a violation of OpenAI's terms of service if proven true.
Alex Tatham: I am certain there will be price reductions for Microsoft and ChatGPT to compete.
Katie Sloan: Vendors will need to differentiate by either doubling down on their proprietary AI capabilities, or adopting a more open, collaborative approach to AI development.
We could see more strategic partnerships, increased focus on custom AI models tailored for enterprise use, and enhanced AI-driven cloud services.
In the UK channel, businesses should be prepared for a wave of marketing narratives and product positioning from vendors vying to stay ahead in the AI race.
Eric Mink: The hyperscaler vendors may shift toward bundling their AI offerings into, for example, Windows instead of selling it as a standalone SKU.
For example, Microsoft has already taken this approach with Microsoft 365, now renamed Microsoft 365 Copilot.
CRN: Will this reduce the value of GenAI add-on overall?
Mini Biswas: Microsoft Copilot continues to make waves with its integration into Microsoft 365 apps.
This is their major differentiator amongst other things such as data security and it being within a customer’s tenant, making work more secure, efficient and productive.
With the recent announcement of a free plan for GitHub Copilot, Microsoft is also ensuring that AI tools are accessible to a wider audience.
In short, no, there is a ton of value of GenAI add-on overall.
As models such as DeepSeek AI are significantly more cost-effective in terms of overall cost of implementation, this may lead to vendors exploring consumption-based models for their customers and making it visible of what their standard and premium features include.
Alex Tatham: No. We have hardly scratched the surface so far.
The opportunity is big enough for all to take advantage.
Perhaps the projections might be reduced - but they were so big anyway that we won't notice.
Katie Sloan: Not necessarily.
While DeepSeek might push AI innovation forward, the value of GenAI add-ons will always depend on their practical application.
Businesses still require tailored, industry-specific AI solutions rather than generic AI capabilities.
If anything, this development could make GenAI even more competitive, encouraging vendors to enhance differentiation through unique business use cases, integrations, and improved cost efficiencies.
Eric Mink: Yes, the current high-priced add-ons we are seeing today are unsustainable if competitors offer similar capabilities for a fraction of the price.
I'm really anticipating price drops, bundling strategies, and a shift in AI pricing models from all current players.
CRN: Is this the sign of a paradigm shift in terms of AI from NORAM to APAC? Should the channel strengthen its focus on Asia?
Mini Biswas: The rise of AI in the APAC region does indicate a potential paradigm shift, and it would be prudent for businesses to strengthen their focus on Asia to leverage the opportunities this shift presents.
However, a few red flags to take into consideration are the geopolitical implications.
Given DeepSeek's Chinese origins, there is likelihood of this affecting the adoption and integration into Western markets.
There is also a red flag on the AI ethics, responsible AI measures we have within the UK, EU versus APAC as well as data sovereignty.
As we start to enter a growing AI market, it is vital to gain insight into AI practices, transparency, fairness, accountability and also learn more about the development into these AI tools.
I personally couldn't easily find any specific information whether there was inclusion of women or diverse backgrounds in DeepSeek's AI dev team, would welcome more news around this.
Katie Sloan: Whilst my knowledge and experience is predominantly in the use of AI by businesses in the UK, I believe this could potentially mark a more pronounced shift in AI leadership towards APAC, given the region’s growing influence in AI research, semiconductor/chip production, and enterprise adoption.
However, the UK channel’s focus should remain on how these advancements translate into real-world business impact.
Whether innovation comes from the US, China, or elsewhere, the role of IT providers and MSPs remains the same: to help businesses navigate AI adoption in a way that drives measurable outcomes.
That said, forward-thinking partners may find opportunities in strengthening ties with APAC vendors and ecosystems, to access emerging AI capabilities sooner.
Ultimately, while DeepSeek introduces new dynamics, the principles of scalable business growth remain unchanged - leveraging technology with a clear strategy, ensuring governance, and aligning IT investments with business objectives.
The channel's role in guiding this transition will be more critical than ever.
Eric Mink: Overall, the channel and end customers should benefit from increasing competitiveness, efficiency gains and, ultimately, cost reduction based on all of AI's current development.
Alex Tatham: Never under-estimate China and APAC - but I can't see Western businesses queuing up to use servers outside North America and Europe.
If the US won't let citizens watch people dancing on TikTok because they are scared of Chinese surveillance - Western organisations and governments are not going to share very sensitive data and information.
However, we are all dancing to DeepSeek's tune now.