Key Takeaways:
- The potential imposition of tariffs in the United States on goods from various countries is – like a lot of things from the second Trump administration – uncertain at this early stage, but the threat of widespread upheaval to established trading routes and agreements is likely to push fashion brands even further towards supply chain diversification and agility. Technology may have some, but not all, of the answers in tools like risk management, supply chain virtualisation and simulation, AI, and on-demand production.
- A new company, Haut.AI, is promoting a system designed for beauty brands and consumers, which uses AI to replicate the effects of skincare products and treatments on different kinds of skin. This kind of tech-enabled personalisation has potentially broad applicability and holds the promise of a measurable return on investment for a technology category that is still struggling to transition from speculative to quantifiable value.
If the US tariffs do come, what tools will fashion brands need to adapt to a suddenly-transformed sourcing landscape?
As a candidate, Donald Trump advocated for broad tariffs, suggesting duties of up to 20% on imports from all countries, a 25% tax on goods from Mexico and Canada, and a 60% tariff on products from China. Now, freshly inaugurated, this week Trump has warned that he is considering a 10% tariff on all Chinese goods, starting as early as the 1st of February. The EU has now also been included, with tariffs being threatened so as to ensure “fairness”.
Among other political reasons, the idea seems to be that steep tariffs would incentivise companies (fashion brands included) to move production back to the United States, away from their international partners. But is the US ready for this kind of production? Shoring up domestic sourcing and manufacturing, after all, presupposes that the right knowledge, capabilities, and infrastructure exist, at the right scale, to offset any scaling-back of the long-prevailing offshore model.
There is definitely a case to be made for US reshoring, but at the moment, the US lacks the skills and technology that’s become so heavily concentrated in the foreign manufacturing hubs that brands of all sizes rely on. And replicating (or leap-frogging) that is not something that can be done overnight.
In the meantime, whether these tariffs end up being implemented, watered down, or abandoned, the message is clear: brands that don’t want to be left holding the proverbial bag will want to diversify their sourcing and manufacturing bases as much as possible, and will also need to ensure that they are as agile as they can be – a requirement for any successful rapid shifting of procurement.
Let’s consider the calculus for brands diversifying away from regions where the US administration has imposed high tariffs. On the positive side, these new jurisdictions with lower or no tariffs could help reduce both brands’ own costs and the impact on consumers, since tariffs are likely to wind up being partially absorbed by the importing company and otherwise passed on to the consumer. But new sourcing regions may still be more expensive in terms of absolute cost of materials, labour etc, potentially affecting the affordability brands currently enjoy with their established partners.
Then there is the human element: many brands have spent years cultivating strong relationships with their overseas suppliers and manufacturers, fine-tuning operations despite language barriers and tight deadlines. Disrupting these partnerships could introduce new challenges in maintaining quality, efficiency, and reliability. Lastly, there is the skills and technology aspect – which is a big one. For example, certain categories – particularly intricate and heavily embellished products like occasionwear – are where China truly excels, as reported by BoF. And, of course, regions outside of China also rely quite heavily on their textile and garment export trade for overall prosperity, and it will not be the fault of those garment workers if those countries fall under the aegis of new tariffs and brands are forced to withdraw their business, shrinking those economies in the process.
There are no easy answers to the unwinding of globalisation, any more than there were to why the conditions that created it grew unchecked for so long. But in the absence of those answers, brands realistically have little choice other than to prioritise making their supply chains more agile and responsive – with technology being the most effective way to achieve this. As a starting point, brands can, of course, make sure that they have robust, real-time visibility into their entire supply chain – from raw materials to finished goods. There are clearly also other arguments for doing this (legislation foremost among them). Better communication with suppliers can help with decision-making, especially if needing to be done quickly, and ensure that production plans can be adjusted quickly in response to tariffs or other disruptions.
Then there is the option for brands to move towards making use of digital platforms that support flexible manufacturing and allow them to quickly switch between different suppliers or production methods, depending on tariff changes or other external factors. Paired with on-the-ground investment in microfactories and larger facilities that are similarly built to make short-run production cost-effective, these platforms can also facilitate on-demand production and rapid prototyping, allowing fashion brands to stay competitive even when supply chain conditions fluctuate.
Last, there is AI. The possibilities for AI upstream are vast and largely unproven, but some examples of what it could be used for include optimising material planning, allocating inventory, predicting production timelines, identifying and modelling risks, and helping brands assess the impact of tariffs on cost structures and adjust pricing strategies accordingly.
This is a tenuous link right now, but in theory some of that positive impact of AI could come from the same source as the tariffs themselves, with the new US government’s The Stargate Project, described as “the largest AI infrastructure project, by far, in history.” Whether this project is realistic is, of course, an open question, but to give us an idea of scale we’re talking about the creator of ChatGPT, OpenAI, teaming up with US tech company Oracle, Japan’s Softbank, and MGX, a tech investment arm of the United Arab Emirates government to build $500bn of AI infrastructure on the ground, in the United States. Other technology partners include British chipmaker Arm, US chipmaker Nvidia, and Microsoft, which already has a tens-of-billions partnership with OpenAI.
According to a 2024 McKinsey report called “AI power: Expanding data center capacity to meet growing demand” global demand for data center capacity is expected to more than triple by 2030, with an annual growth rate between 19% and 27%. To meet this demand, McKinsey predicts that developers will need to construct at least twice the capacity by 2030 compared to what has been built since 2000. Oracle CEO Larry Ellison mentioned during a briefing that 10 data centers for the project are already under construction in Texas, with more planned for the future.
While fashion and other industries are looking at the near-term impact of reshoring and rebalancing their supply chains, there is at least the possibility that the AI being held out as the solution to the problem could wind up running its inference in-country.
SkinGPT: the new era for beauty of show, don’t tell
Earlier this week, Haut.AI – a company created to sell AI-driven beauty intelligence solutions – announced the commercial launch of SkinGPT, a generative AI platform designed to revolutionise virtual skincare try-ons in the beauty industry. Tailored for beauty brands, SkinGPT uses AI to virtualise the effects of skincare products and treatments, offering personalised, accurate, and interactive virtual experiences that will improve the way consumers discover and buy skincare products.
According to the company, SkinGPT integrates with e-commerce platforms, allowing for virtual product try-ons directly on product pages. Marketers can use the tool to create visuals, while researchers can use the same software in the formulation and creation of the actual products. And Haut.AI has some big names attached to it already: the company is backed by Ulta Beauty, a member of Nvidia’s start-up programme, and whose parserships include Clarins and Unilever.
The same company is also launching Generative Skin for consumers, allowing them to explore how skincare ingredients impact their skin. Users upload a photo to receive what the company calls “realistic” visualisations of how ingredients like retinol or Vitamin C might address concerns such as pigmentation, redness, and breakouts.
There is, we should note, the long-running question of whether generative AI outputs at the image and video level can be called “simulations”. The Interline has dedicated several articles to this idea in the past.
In an interview, Estonia-based founder Anastasia Georgievskaya said that the mission behind the company, initially, was to find differences in skin conditions across demographics, lifestyles, and geographical locations. Georgievskaya describes how Haut.AI’s software was built on years of clinical development for Contract Research Organizations (CROs), designed to precisely measure before-and-after effects.
As per her latest research, Georgievskaya applied AI and computer vision to analyse over 17,000 selfies, identifying 136 skin attributes and mapping variations in skin aging biomarkers across different demographics. The study highlighted how factors like gender, age, and lifestyle significantly influence skin standards, and it spotlighted how a lack of individualised benchmarks has caused unrealistic, broad-brush consumer expectations that the beauty industry has been struggling to meet.
How far beauty consumers will trust in these visualisations remains to be seen, but this is nevertheless further evidence of just how far the combination of accessible hardware (the humble smartphone) paired with advanced analytical and generative models could transform the buying experience for one of the industry’s fastest-growing segments.
Best from The Interline:
This week saw publication of the second in our ‘DPC Designer’ series. Yasmin Koppe created for the “Apparel” category in our DPC Report 2024. We sat down with Yasmin to cover everything from creative inspiration and inclusivity, to hardware, pipelines, and cross-industry best practices.
And we wrapped up the series with Canberk Karakas, who created for the “Footwear” category in our DPC Report 2024. We chatted to Canberk about how he approached the challenge of designing for The Interline, how he believes the footwear industry is embracing 3D, and what’s next for DPC in education and industry use cases.